212 TJie American Geologist. April, 1900 
amounts of the constituent minerals ; and the succession of 
groups should be arranged as nearly as possible according to 
the known laws of mineral associations, so as to form a con- 
nected series. It may be advisable, also, in the end, to en- 
large the number of structural variations or types by the sep- 
aration of each of the first two types given into two, on the 
basis of fine or coarse grain, these typ^s to be distinguished 
by appropriate suffixes. Thus there would be a coarse granu- 
lar hypidiomorphic type and a fine granular hypidiomorphic 
type, a coarse panidiomorphic type and a fine-grained pani- 
diomorphic type. 
Rocks not belonging to the feldspar class, such as the 
nepheline and leucite rocks, should be classified according to 
the same principles as the feldspar rocks, and intermediate 
classes should cover the transitions. Between feldspar rocks, 
and the nepheline and leucite rocks, for example, is the com- 
mon transitional class in which occur eleolite- syenyte and 
phonolyte. Under each of the classes the rocks should be 
classified according to family, group, species, variety and type, 
but without too much analogy to the feldspar rocks implied 
in their names. A single example of this false analogy is the 
leucite-basalt. This is chemically the equivalent of a syenyte'" 
and the classing and naming of it with the basalts is a mistake, 
originating in the idea that olivine was confined to the basalts 
among feldspar rocks and that it might serve as a prime 
element in classification. Olivine may occur in monzonytest 
and in very acid andesytes;];, and is certainly not confined to the 
basalts as at present understood, so cannot be made a basis 
for classification of such rocks as leucite-basalt. Character- 
istic names, suggesting no misleading analogies, are necessary. 
Different classes of rocks should be represented in the 
table side by side, arranged in their natural order, and the dif- 
ferent groups in all classes other than the feldspar class should 
be, whenever possible, so arranged that each will fall laterally 
opposite the position of the chemically equivalent group in the 
*Principes a suivre pour une Classification universelle des Roches. 
Par M. A. Michel-Levy. Compte-Rendu. Lausanne. Troisieme par- 
tie, p. 152. Bull. U. S. Geol. Survey, Xo. 148. p. 153. 
•tBrogger: Triadische Eruptionsfolge bei Predazzo. Kristiania 1895. 
p. 47- 
JM. A. Michel-L6vy, op. cit , p. 156. 
